Hello trend,
This is Satya Mallick from LearnOpenCV.com.
This week you will learn about Multiple Object Tracking and Re-Identification using FairMOT. This post is part of our popular Object Tracking series.
What is Multiple Object Tracking and Re-ID?
Multiple Object Tracking is the process of finding multiple objects and keeping track of their trajectories in a video sequence.
In case of Object Detection, it is enough to know the bounding boxes of each detected object, but in some applications such as surveillance, it is important to track the identity of a target object. Imagine automatically identifying a criminal across multiple cameras as they try to hide or escape!
The process of preserving the identities of the detected objects along with their trajectories is called Re-Identification.
What is FairMOT
FairMOT performs object detection and re-identification for Multiple Object Tracking using a single network. In today's post, we understand how the algorithm works, and how to perform object tracking using FairMOT.
FairMOT is fast and accurate! It was released in April 2020 and is among the most popular methods for Multiple Object Tracking according to paperswithcode data.
Without further ado, let's get started.
| Object Tracking and Re-Id with FairMOT |
You can check out the code by clicking on the button below, and star us on GitHub to say thanks!
| Download Code (GitHub) |
Highlights from the post
Check out the 3-minute video covering the highlights of the blog post:
P.S. If you have come across an awesome Object Tracking or Re-ID model, please let us know and we will research and write a post on the same.
NVIDIA GTC Sneak Peek + How to Win a 3080Ti?
There has been a lot of buzz around NVIDIA GTC 2022 - the latest conference from NVIDIA on everything related to AI! Here's a sneak peek video on the same.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.